This is not an unrecoverable situation. Users of kvm_read_guest_offset_cached
and kvm_write_guest_offset_cached must expect the read/write to fail, and
therefore it is possible to just return early with an error value.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reject userspace memslots whose size exceeds the storage capacity of an
"unsigned long". KVM's uAPI takes the size as u64 to support large slots
on 64-bit hosts, but does not account for the size being truncated on
32-bit hosts in various flows. The access_ok() check on the userspace
virtual address in particular casts the size to "unsigned long" and will
check the wrong number of bytes.
KVM doesn't actually support slots whose size doesn't fit in an "unsigned
long", e.g. KVM's internal kvm_memory_slot.npages is an "unsigned long",
not a "u64", and misc arch specific code follows that behavior.
Fixes: fa3d315a4c ("KVM: Validate userspace_addr of memslot when registered")
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Message-Id: <20211104002531.1176691-3-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When modifying memslots, snapshot the "old" memslot and copy it to the
"new" memslot's arch data after (re)acquiring slots_arch_lock. x86 can
change a memslot's arch data while memslot updates are in-progress so
long as it holds slots_arch_lock, thus snapshotting a memslot without
holding the lock can result in the consumption of stale data.
Fixes: b10a038e84 ("KVM: mmu: Add slots_arch_lock for memslot arch fields")
Cc: stable@vger.kernel.org
Cc: Ben Gardon <bgardon@google.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20211104002531.1176691-2-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
* Fixes for Xen emulation
* Kill kvm_map_gfn() / kvm_unmap_gfn() and broken gfn_to_pfn_cache
* Fixes for migration of 32-bit nested guests on 64-bit hypervisor
* Compilation fixes
* More SEV cleanups
In commit 7e2175ebd6 ("KVM: x86: Fix recording of guest steal time /
preempted status") I removed the only user of these functions because
it was basically impossible to use them safely.
There are two stages to the GFN->PFN mapping; first through the KVM
memslots to a userspace HVA and then through the page tables to
translate that HVA to an underlying PFN. Invalidations of the former
were being handled correctly, but no attempt was made to use the MMU
notifiers to invalidate the cache when the HVA->GFN mapping changed.
As a prelude to reinventing the gfn_to_pfn_cache with more usable
semantics, rip it out entirely and untangle the implementation of
the unsafe kvm_vcpu_map()/kvm_vcpu_unmap() functions from it.
All current users of kvm_vcpu_map() also look broken right now, and
will be dealt with separately. They broadly fall into two classes:
* Those which map, access the data and immediately unmap. This is
mostly gratuitous and could just as well use the existing user
HVA, and could probably benefit from a gfn_to_hva_cache as they
do so.
* Those which keep the mapping around for a longer time, perhaps
even using the PFN directly from the guest. These will need to
be converted to the new gfn_to_pfn_cache and then kvm_vcpu_map()
can be removed too.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Message-Id: <20211115165030.7422-8-dwmw2@infradead.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Generalize KVM_REQ_VM_BUGGED so that it can be called even in cases
where it is by design that the VM cannot be operated upon. In this
case any KVM_BUG_ON should still warn, so introduce a new flag
kvm->vm_dead that is separate from kvm->vm_bugged.
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
All of the irqfds would to be updated when update the irq
routing, it's too expensive if there're too many irqfds.
However we can reduce the cost by avoid some unnecessary
updates. For irqs of MSI type on X86, the update can be
saved if the msi values are not change.
The vfio migration could receives benefit from this optimi-
zaiton. The test VM has 128 vcpus and 8 VF (with 65 vectors
enabled), so the VM has more than 520 irqfds. We mesure the
cost of the vfio_msix_enable (in QEMU, it would set routing
for each irqfd) for each VF, and we can see the total cost
can be significantly reduced.
Origin Apply this Patch
1st 8 4
2nd 15 5
3rd 22 6
4th 24 6
5th 36 7
6th 44 7
7th 51 8
8th 58 8
Total 258ms 51ms
We're also tring to optimize the QEMU part [1], but it's still
worth to optimize the KVM to gain more benefits.
[1] https://lists.gnu.org/archive/html/qemu-devel/2021-08/msg04215.html
Signed-off-by: Longpeng(Mike) <longpeng2@huawei.com>
Message-Id: <20210827080003.2689-1-longpeng2@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
KVM_MAX_VCPU_ID is not specifying the highest allowed vcpu-id, but the
number of allowed vcpu-ids. This has already led to confusion, so
rename KVM_MAX_VCPU_ID to KVM_MAX_VCPU_IDS to make its semantics more
clear
Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210913135745.13944-3-jgross@suse.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm_make_vcpus_request_mask() already disables preemption so just like
kvm_make_all_cpus_request_except() it can be switched to using
pre-allocated per-cpu cpumasks. This allows for improvements for both
users of the function: in Hyper-V emulation code 'tlb_flush' can now be
dropped from 'struct kvm_vcpu_hv' and kvm_make_scan_ioapic_request_mask()
gets rid of dynamic allocation.
cpumask_available() checks in kvm_make_vcpu_request() and
kvm_kick_many_cpus() can now be dropped as they checks for an impossible
condition: kvm_init() makes sure per-cpu masks are allocated.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210903075141.403071-9-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Allocating cpumask dynamically in zalloc_cpumask_var() is not ideal.
Allocation is somewhat slow and can (in theory and when CPUMASK_OFFSTACK)
fail. kvm_make_all_cpus_request_except() already disables preemption so
we can use pre-allocated per-cpu cpumasks instead.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210903075141.403071-8-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Iterating over set bits in 'vcpu_bitmap' should be faster than going
through all vCPUs, especially when just a few bits are set.
Drop kvm_make_vcpus_request_mask() call from kvm_make_all_cpus_request_except()
to avoid handling the special case when 'vcpu_bitmap' is NULL, move the
code to kvm_make_all_cpus_request_except() itself.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210903075141.403071-5-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Check for a NULL cpumask_var_t when kicking multiple vCPUs via
cpumask_available(), which performs a !NULL check if and only if cpumasks
are configured to be allocated off-stack. This is a meaningless
optimization, e.g. avoids a TEST+Jcc and TEST+CMOV on x86, but more
importantly helps document that the NULL check is necessary even though
all callers pass in a local variable.
No functional change intended.
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20210827092516.1027264-3-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Fix a benign data race reported by syzbot+KCSAN[*] by ensuring vcpu->cpu
is read exactly once, and by ensuring the vCPU is booted from guest mode
if kvm_arch_vcpu_should_kick() returns true. Fix a similar race in
kvm_make_vcpus_request_mask() by ensuring the vCPU is interrupted if
kvm_request_needs_ipi() returns true.
Reading vcpu->cpu before vcpu->mode (via kvm_arch_vcpu_should_kick() or
kvm_request_needs_ipi()) means the target vCPU could get migrated (change
vcpu->cpu) and enter !OUTSIDE_GUEST_MODE between reading vcpu->cpud and
reading vcpu->mode. If that happens, the kick/IPI will be sent to the
old pCPU, not the new pCPU that is now running the vCPU or reading SPTEs.
Although failing to kick the vCPU is not exactly ideal, practically
speaking it cannot cause a functional issue unless there is also a bug in
the caller, and any such bug would exist regardless of kvm_vcpu_kick()'s
behavior.
The purpose of sending an IPI is purely to get a vCPU into the host (or
out of reading SPTEs) so that the vCPU can recognize a change in state,
e.g. a KVM_REQ_* request. If vCPU's handling of the state change is
required for correctness, KVM must ensure either the vCPU sees the change
before entering the guest, or that the sender sees the vCPU as running in
guest mode. All architectures handle this by (a) sending the request
before calling kvm_vcpu_kick() and (b) checking for requests _after_
setting vcpu->mode.
x86's READING_SHADOW_PAGE_TABLES has similar requirements; KVM needs to
ensure it kicks and waits for vCPUs that started reading SPTEs _before_
MMU changes were finalized, but any vCPU that starts reading after MMU
changes were finalized will see the new state and can continue on
uninterrupted.
For uses of kvm_vcpu_kick() that are not paired with a KVM_REQ_*, e.g.
x86's kvm_arch_sync_dirty_log(), the order of the kick must not be relied
upon for functional correctness, e.g. in the dirty log case, userspace
cannot assume it has a 100% complete log if vCPUs are still running.
All that said, eliminate the benign race since the cost of doing so is an
"extra" atomic cmpxchg() in the case where the target vCPU is loaded by
the current pCPU or is not loaded at all. I.e. the kick will be skipped
due to kvm_vcpu_exiting_guest_mode() seeing a compatible vcpu->mode as
opposed to the kick being skipped because of the cpu checks.
Keep the "cpu != me" checks even though they appear useless/impossible at
first glance. x86 processes guest IPI writes in a fast path that runs in
IN_GUEST_MODE, i.e. can call kvm_vcpu_kick() from IN_GUEST_MODE. And
calling kvm_vm_bugged()->kvm_make_vcpus_request_mask() from IN_GUEST or
READING_SHADOW_PAGE_TABLES is perfectly reasonable.
Note, a race with the cpu_online() check in kvm_vcpu_kick() likely
persists, e.g. the vCPU could exit guest mode and get offlined between
the cpu_online() check and the sending of smp_send_reschedule(). But,
the online check appears to exist only to avoid a WARN in x86's
native_smp_send_reschedule() that fires if the target CPU is not online.
The reschedule WARN exists because CPU offlining takes the CPU out of the
scheduling pool, i.e. the WARN is intended to detect the case where the
kernel attempts to schedule a task on an offline CPU. The actual sending
of the IPI is a non-issue as at worst it will simpy be dropped on the
floor. In other words, KVM's usurping of the reschedule IPI could
theoretically trigger a WARN if the stars align, but there will be no
loss of functionality.
[*] https://syzkaller.appspot.com/bug?extid=cd4154e502f43f10808a
Cc: Venkatesh Srinivas <venkateshs@google.com>
Cc: Vitaly Kuznetsov <vkuznets@redhat.com>
Fixes: 97222cc831 ("KVM: Emulate local APIC in kernel")
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20210827092516.1027264-2-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
grow_halt_poll_ns() ignores values between 0 and
halt_poll_ns_grow_start (10000 by default). However,
when we shrink halt_poll_ns we may fall way below
halt_poll_ns_grow_start and endup with halt_poll_ns
values that don't make a lot of sense: like 1 or 9,
or 19.
VCPU1 trace (halt_poll_ns_shrink equals 2):
VCPU1 grow 10000
VCPU1 shrink 5000
VCPU1 shrink 2500
VCPU1 shrink 1250
VCPU1 shrink 625
VCPU1 shrink 312
VCPU1 shrink 156
VCPU1 shrink 78
VCPU1 shrink 39
VCPU1 shrink 19
VCPU1 shrink 9
VCPU1 shrink 4
Mirror what grow_halt_poll_ns() does and set halt_poll_ns
to 0 as soon as new shrink-ed halt_poll_ns value falls
below halt_poll_ns_grow_start.
Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210902031100.252080-1-senozhatsky@chromium.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
KVM/arm64 updates for 5.15
- Page ownership tracking between host EL1 and EL2
- Rely on userspace page tables to create large stage-2 mappings
- Fix incompatibility between pKVM and kmemleak
- Fix the PMU reset state, and improve the performance of the virtual PMU
- Move over to the generic KVM entry code
- Address PSCI reset issues w.r.t. save/restore
- Preliminary rework for the upcoming pKVM fixed feature
- A bunch of MM cleanups
- a vGIC fix for timer spurious interrupts
- Various cleanups
Add a new stat that counts the number of times a remote TLB flush is
requested, regardless of whether it kicks vCPUs out of guest mode. This
allows us to look at how often flushes are initiated.
Unlike remote_tlb_flush, this one applies to ARM's instruction-set-based
TLB flush implementation, so apply it there too.
Original-by: David Matlack <dmatlack@google.com>
Signed-off-by: Jing Zhang <jingzhangos@google.com>
Message-Id: <20210817002639.3856694-1-jingzhangos@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Don't export KVM's MMU notifier count helpers, under no circumstance
should any downstream module, including x86's vendor code, have a
legitimate reason to piggyback KVM's MMU notifier logic. E.g in the x86
case, only KVM's MMU should be elevating the notifier count, and that
code is always built into the core kvm.ko module.
Fixes: edb298c663 ("KVM: x86/mmu: bump mmu notifier count in kvm_zap_gfn_range")
Cc: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-Id: <20210902175951.1387989-1-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add three log histogram stats to record the distribution of time spent
on successful polling, failed polling and VCPU wait.
halt_poll_success_hist: Distribution of spent time for a successful poll.
halt_poll_fail_hist: Distribution of spent time for a failed poll.
halt_wait_hist: Distribution of time a VCPU has spent on waiting.
Signed-off-by: Jing Zhang <jingzhangos@google.com>
Message-Id: <20210802165633.1866976-6-jingzhangos@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Allow archs to create arch-specific nodes under kvm->debugfs_dentry directory
besides the stats fields. The new interface kvm_arch_create_vm_debugfs() is
defined but not yet used. It's called after kvm->debugfs_dentry is created, so
it can be referenced directly in kvm_arch_create_vm_debugfs(). Arch should
define their own versions when they want to create extra debugfs nodes.
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20210730220455.26054-2-peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
